Scoring methods in MALDI peptide mass fingerprinting: ChemScore, and the ChemApplex program

No universally accepted score is currently available to determine when a matrix-assisted laser desorption ionization (MALDI) peptide mass fingerprint (PMF) experiment has been successfully carried out. We describe a software program (ChemApplex) based on a calculated parameter (Combined Protein Score) that takes into account (1) peak intensity, (2) the mass accuracy of the match, and (3) ChemScore, a theoretical intensity factor that estimates the probability of observing a particular peptide based on a combination of chemical considerations, in particular the amino acid composition of the peptide and the amino acid sequence of the amino acids that span the cleavage site. When these three factors are taken into account both at the level of individual peptides and at the protein level, protein components in mixtures whose peptides contribute less than 1% of the total intensity can often be correctly identified, as is demonstrated for mixtures of standard proteins. Moreover, it is possible to make robust database identifications that are nearly independent of the number of masses submitted and the mass error threshold used for matching. Protein scoring based on Combined Protein Score is orthogonal to many of the commonly used probability-based scoring schemes, and makes it possible to archive a more complete set of parameters that more thoroughly characterize the validity of the database match, which increases the confidence in the identifications. (J Am Soc Mass Spectrom 2002, 13, 22-39) (C) 2002 American Society for Mass Spectrometry.